cialyoucanletgo

BioMedCentralPage1of9(pagenumbernotforcitationpurposes)HealthandQualityofLifeOutcomesOpenAccessResearchPsychologicalwellbeing,physicalimpairmentsandruralaginginadevelopingcountrysettingMelanieAAbas*1,SureepornPunpuing2,TawanchaiJirapramupitak3,KanchanaTangchonlatip2andMorvenLeese1Address:1HealthServiceandPopulationResearchDepartment,King'sCollegeLondon,London,UK,2InstituteofPopulationandSocialResearch,MahidolUniversity,Nakhonpathom,Thailandand3FacultyofMedicine,ThammasatUniversity,Pathumthani,ThailandEmail:MelanieAAbas*-m.
abas@iop.
kcl.
ac.
uk;SureepornPunpuing-prspu@mahidol.
ac.
th;TawanchaiJirapramupitak-tawanchaij@gmail.
com;KanchanaTangchonlatip-prktc@mahidol.
ac.
th;MorvenLeese-M.
Leese@iop.
kcl.
ac.
uk*CorrespondingauthorEqualcontributorsAbstractBackground:Therehasbeenverylittleresearchonwellbeing,physicalimpairmentsanddisabilityinolderpeopleindevelopingcountries.
Methods:Acommunitysurveyof1147olderparents,oneperhousehold,agedsixtyandoverinruralThailand.
WeusedtheBurvillscaleofphysicalimpairment,theThaiPsychologicalWellbeingScaleandthebriefWHODisabilityAssessmentSchedule.
Weratedreceivedandperceivedsocialsupportseparatelyfromchildrenandfromothersandratedsupporttochildren.
Weusedweightedanalysestotakeaccountofthesamplingdesign.
Results:Impairmentsduetoarthritis,pain,paralysis,vision,stomachproblemsorbreathingwereallassociatedwithlowerwellbeing.
Afteradjustingfordisability,onlyimpairmentduetoparalysiswasindependentlyassociatedwithloweredwellbeing.
Theeffectofhavingtwoormoreimpairmentscomparedtononewasassociatedwithloweredwellbeingafteradjustingfordemographicfactorsandsocialsupport(adjusteddifference-2.
37onthewell-beingscalewithSD=7.
9,p<0.
001)butafteradjustingfordisabilitythecoefficientfellandwasnon-significant.
Theparsimoniousmodelforwellbeingincludedage,wealth,socialsupport,disabilityandimpairmentduetoparalysis(theeffectofparalysiswas-2.
97,p=0.
001).
InthisThaisetting,receivedsupportfromchildrenandfromothersandperceivedgoodsupportfromandtochildrenwereallindependentlyassociatedwithgreaterwellbeingwhereasactualsupporttochildrenwasassociatedwithlowerwellbeing.
Lowreceivedsupportfromchildreninteractedwithparalysisinbeingespeciallyassociatedwithlowwellbeing.
Conclusion:InthisThaisetting,asfoundinwesternsettings,mostoftheassociationbetweenphysicalimpairmentsandlowerwellbeingisexplainedbydisability.
Disabilityispotentiallymediatingtheassociationbetweenimpairmentandlowwellbeing.
Receivedsupportmaybuffertheimpactofsomeimpairmentsonwellbeinginthissetting.
Givingactualsupporttochildrenisassociatedwithlesswellbeingunlessthesupportbeinggiventochildrenisperceivedasgood,perhapsreflectingparentalobligationtosupportadultchildreninneed.
Improvingcommunitydisabilityservicesforolderpeopleandoptimizingreceivedsocialsupportwillbevitalinruralareasindevelopingcountries.
Published:16July2009HealthandQualityofLifeOutcomes2009,7:66doi:10.
1186/1477-7525-7-66Received:2March2009Accepted:16July2009Thisarticleisavailablefrom:http://www.
hqlo.
com/content/7/1/662009Abasetal;licenseeBioMedCentralLtd.
ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense(http://creativecommons.
org/licenses/by/2.
0),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
HealthandQualityofLifeOutcomes2009,7:66http://www.
hqlo.
com/content/7/1/66Page2of9(pagenumbernotforcitationpurposes)BackgroundThereisincreasinginterestworldwideinthestudyofwell-beingasameanstoassessneedandtoevaluatepositivedimensionsofhealthcareprograms.
Positivementalhealth"whichallowsindividualstorealisetheirabilities,cope,andcontributetotheircommunities"[1]andthecapacitytosustainsocialrelationshipsarekeydimensionsofwellbeing[2].
Wellbeingcanbemeasuredintermsofpositivepsychologicalsymptoms(suchasbeingabletoenjoythingsandtoletgoofworries)orlifesatisfaction,butincreasinglymultidimensionalscalesareusedwhichincludeconceptssuchasautonomy,self-acceptanceandrelationswithothers[3,4].
Researchonassociationsbetweenphysicalimpairmentsandwellbeinginolderpeoplehasbeenlimited[5-7]althoughtherehavebeenseveralstudiesofdepressionasanoutcomesuggestingthatdisabilitymediatesmostoftheeffectofspecificmedicalconditionsondepression[8-10].
However,researchuntilnowhascomealmostentirelyfromricherindustrialisedcountries.
Oneaimofthisstudywastoseewhetherpatternsofassociationbetweenimpairment,disabilityandpsychologicalwell-beinginThailandaresimilartoordifferentfromthosedescribedelsewhere.
Givencross-culturaldifferencesinperceivedwell-being,arecentadvancehasbeentodevelopculture-specificscalessuchastheChineseAgingWellProfile(2007)[11].
InThailand,Ingersoll-Daytonetal[12]developedandvalidatedtheThaipsychologicalwell-beingscale,whichisrelatedtotheScaleofPsycho-logicalWell-beingScale[3].
Particularfeaturesofthis,whichistheonlymultidimensionalwellbeingscaledevel-opedforusewithThaiolderpeople,isthatcomparedtoversionsusedinWesternsettings,moreofthedimensionsareinterpersonal(measuringharmonyandinterconnect-ednesswithotherpeople)andfewerareintrapersonal(e.
g.
measuringacceptanceandpositivemood).
InThailand,thesettingforthisstudy,theproportionofadults60yearsofageandoverrosefrom4.
5%in1960to9.
5%in2000andispredictedtobe25%in2040[13].
IntheruralThaicontext,asinmanydevelopingcountries,facilitiesforhealthcareandsupportfordisabilitiesarelimited.
Alsoinmanyotherdevelopingcountries,rapidriseinruraltourbanmigrationofyoungadultsmeansthatolderparentsareincreasinglylivingseparatelyfromtheiradultchildren[14].
InThailandasinotherAsiancul-tures,childrentraditionallytakeresponsibilityforolderparentsandolderparentscontinuetosupportchildren.
Giventhepotentialrelativeimportanceofsupportfromchildren[15]wewereinterestedtoseeifsupportfromchildrenratherthansupportfromotherswasassociatedwithwellbeing.
MethodsSettingWenestedthestudywithintheKanchanaburiDemo-graphicSurveillanceSysteminwesternThailand[16].
Kanchanaburiprovinceisamostlyruralregionlocated130kilometreswestofBangkokwithapopulationofabout735,000in2007.
TheKanchanaburiDemographicSurveillanceSystemsystemhasmonitoredhouseholdssince2000in100neighborhoods(villagesandurbancen-susblocks).
Theneighborhoodsweredrawnfromfivestrata(classifiedonecological,socio-economicandpopu-lationcriteria)bystratifiedrandomsamplingfromtheprovincepopulationof871villagesand131urbancensusblocks.
Thestudydescribedhereispartofalongitudinalstudydesignedtostudytheimpactonolderparentsofout-migrationoftheiradultchildren/offspring[17]Dur-ingsamplingforthemainstudyweneededtoidentifywhicholderadultswereparentsofatleastonelivingchildoffspring,andwhethertheolderparentwasco-residentornotwithatleastoneoftheiroffspring.
Therewasapoten-tialsampleof3916householdswithatleastoneolderadultaged60andabove,ofwhom2432(62%)hadatleastonechildoffspringoftheolderadultinthesamehousehold,and1484(38%)didnot.
Weusedsimpleran-domsamplingtoselect60%ofhouseholdswhereanolderadultwasnotco-residentwithatleastoneoftheirchildoffspringand30%ofhouseholdswhereanolderadultwasco-residentwithatleastoneoftheirchildoff-spring.
Thiscomprisedatotalof1620households.
Weusedrandomselectiontoidentifytheparticipantinsitua-tionswheretherewasmorethanoneeligibleparentlivinginahousehold.
DatawerecollectedfromNovember2006toJan2007.
RecruitmentTheinterviewingteamvisitedeachsamplingunitandmadecontactwiththevillageheadmanpriortovisitingeachselectedhousehold.
Thepopulationsweremostlyalreadywellacquaintedwiththedemographicsurveil-lancesystem.
Iftheselectedolderadultandthehouseholdheadgaveconsent,theinterviewerfirstinterviewedthehouseholdheadwiththehouseholdquestionnaireandthentheolderadultwiththeindividualquestionnaire.
QuestionnairedevelopmentWecarriedoutfocusgroupdiscussionstoexploreexperi-encesofruralageing,healthandwellbeingandexchangeswithfamilymembers.
Thisinformedthedevelopmentofthequestionnairewhichwaspre-testedbyateamoftenexperiencedinterviewersonthreeseparateoccasions.
Aftereachpre-testwemademodificationsbyconsensus.
Thefinalversionwasback-translatedtoEnglishandcheckedforconsistencybyabilingualpsychiatristandabilingualsocialscientist.
HealthandQualityofLifeOutcomes2009,7:66http://www.
hqlo.
com/content/7/1/66Page3of9(pagenumbernotforcitationpurposes)InclusioncriteriaFluentThai-speaking;aged60orover;parentofatleastonelivingchild(biological,adoptedorstep-child);resi-denceinademographicsurveillancesystemvillagesinceatleast2004.
DependentvariablePsychologicalwell-being.
Weusedthe15-itemThaiwell-beingscale[12,18],developedusingextensivequalitativeandquantitativemethods.
Ithasfivedimensionsofwell-beingwhichareharmony,interdependencewithclosepersons,respect(fromothers),acceptanceandenjoy-ment.
Eachdimensionhasthreeitemswhichweredevel-opedfromconfirmatoryfactoranalysis.
WeusedtheglobalfactormodelwhichwasshowninThailandtohavegoodfitindices(goodnessoffit0.
95,rootmeansquareerrorofapproximation0.
05)[12].
Theitemsofthescalehavebeenshowntohaveadequateinternalconsistency(Cronbach'salphacoefficientinthissample0.
89)andtest-retestreliability(rangingfrom0.
6to0.
7inpreviouswork)[12]andthescalecorrelatedpositivelywithlifesat-isfactionandnegativelywiththeGeriatricDepressionScale(-0.
4)[12].
Astatementisreadoutforeachitem.
Forexample,foracceptancethestatementis'Whenyouhavesmallproblems,youcanletgoofyourworries'.
Theolderpersonindicatesona4-pointscaleifthestatementisnotatalltrue,slightlytrue,somewhattrueorverytrue.
IndependentvariablePhysicalIllnessesandImpairments:weusedamodifiedversionoftheBurvillphysicalillnessscale[19].
Partici-pantswereaskedaboutthepresenceof13commonmed-icalproblemsincludingbreathlessness,faints/blackouts,arthritis,paralysis/lossoflimb,skindisorders,hearingdif-ficulties,hearttrouble,eyesightproblems,gastrointestinalproblems,highbloodpressure,diabetesandpain.
Ifanyoftheproblemswaspresentwerateditasimpairmentifparticipantsstatedthattheproblemwasinterferingagreatdealwiththeirfunction.
PotentialconfoundersSocio-economicpositionyearsofeducation,numberofhouseholdassets(outof22,suchasownershipofafridge,motorcycle,ormobilephone),andhouseholdwealthindex.
Weusedprincipalcomponentsanalysistodevelopthehouseholdwealthindexfromthelistofassetsandtheinterviewer'sglobalratingofhouseholdquality.
Thefirstprincipalcompo-nent(whichaccountedfor26%ofthevariancecomparedto7%forthesecondnextmostimportant)wasusedtoprovideanoverallsocioeconomicindexbasedonthese23items.
Thisfinalindexcomprised15items(14house-holdassetsplushouseholdquality).
SocialnetworkandsocialsupportWemodifiedexistingmeasuresinthelightoftheimpor-tanceintheThaicontextofthefamilyandofchildren.
Wemeasuredsizeofneighbourhoodfamilynetwork,fre-quencyoftalkingtoachild,frequencyoftalkingtofriends,receivedsupport(instrumental,emotional,finan-cial),actualsupporttochildren(instrumental,emotional,financial),perceivedadequacyofsupportfromandtochildren,andreceivedsupportfromothers[20-22].
Thereceivedsocialsupportfromchildrenscaleratedreceivedsupportyes/nofromanyoftheirchildrenoneachoftenitems.
Thereceivedsocialsupportfromothersscaleratedreceivedsupportyes/nofromanyoneotherthanchildrenonthesametenitems.
Thesupporttochildrenscaleratedsupporttoanychildrenoneachoffiveitems.
CognitivefunctionweusedalearningtaskwhichhasbeenusedextensivelyinlowandmiddleincomecountrieswhichisdrawnfromtheConsortiumtoEstablishaRegistryofAlzheimer'sDis-ease(CERAD)[23,24],comprisingimmediaterecallanddelayedrecallofaten-wordlist.
Wedefinedsignificantcognitiveimpairmentasperformanceatorbelow1.
5standarddeviationsbelowthenormfortheindividual'sagegroupandeducationallevelonbothtests.
DisabilityWeusedthebrief(12-item)questionnairefromtheWHODisabilityAssessmentScheduletoratedisabilityoverthepast30days[25].
Wewereunabletotranslatetheitemonlearninganewtask,whichwasviewedasnotapplicableforolderadultsinthissetting.
Therefore,weused11items,eachself-ratedonafourpointscalefromnoprob-lemwithcarryingouttheactivitytototal/extremeinabil-ity.
Domainsincludedunderstandingandcommunicatingwiththeworld,gettingaround,self-care,gettingalongwithpeople,activitiesandparticipationinsociety.
Wecategorisedthetotalscoreintothirdsoflow,mediumandhighdisability.
DatacollectionThedatacollectionteamoffoursupervisorsandtwelveinterviewershadatleastabachelor'sdegree.
Mosthadpreviousexperiencewithinterviewingforthedemo-graphicsurveillancesystem.
Residentialtrainingtooktendaysandincludedpresentations,roleplayandpracticeinpilotvillages.
ThestudywaspresentedtotheinterviewersasastudyofhealthyageinginThailand.
Purposefully,nopossiblelinkswerediscussedbetweenpsychologicalwell-being,impairment,disabilityorsocialsupportfromchil-dreninordertoblindtheinterviewerstotheresearchhypothesesandnoneofthesesectionsoftheinterviewimmediatelyfollowedeachotherinsequence.
HealthandQualityofLifeOutcomes2009,7:66http://www.
hqlo.
com/content/7/1/66Page4of9(pagenumbernotforcitationpurposes)Thedatacollectionteamstayedinthevillagesatthehead-man'shouseorthetemple.
Qualitycontrolincludedchecksondatacompletenessandconsistency.
Interview-ershadtoreturntotheparticipantifdatawereinade-quate.
Fieldstationresearchmanagers(trainedintheinterviewbutblindtothehypothesis),andresearcherswereinfrequenttelephonecontactandregularlyvisitedthedatacollectionteams.
WeconductedallinterviewsinThaiandgatheredinformedconsentfromallparticipants.
WegainedethicalapprovalfromKingsCollegeResearchEthicsCommittee(No.
05/05-68)andfromMahidolUni-versityInstitutionalReviewBoard.
SamplesizecalculationThiswasdevelopedforthemainlongitudinalstudywhichwasdesignedtostudytheimpactonolderparentsofout-migrationoftheiradultchildren/offspringfromthedis-trict[17].
Thesamplesizewasbasedonacomparisonofprevalenceofcommonmentaldisorderinthosewithallchildrenmigratedversusthosewithsomechildrenmigratedandrequiredatotalsamplesizeof954giventheproportionsexpectedofthoseexposedandnotexposedtohavingalltheirchildrenmigratefromthedistrict.
AnalysisWeusedStataversion9forWindows(Release9,CollegeStation,TX:StataCorporation.
2003).
Weweightedthedatausingtheproductoftwosetsofprobabilityweightstotakeaccountofdifferentialsamplingatneighbourhoodandhouseholdlevels.
Theweightingatneighbourhoodleveltookaccountoftheprobabilityoftheneighbour-hoodbeingselectedfromthetotalnumberofneighbour-hoodsinthatstratumintheprovince.
Theweightingathouseholdleveltookaccountoftheprobabilityofbeingselectediftheolderparentwasorwasnotco-residentwithoneoftheiroffspring.
WeusedthesurveycommandsinStata(svyset)foranalyses.
Wefirstdescribedtheunad-justedassociationsbetweenwellbeingscoreandthesocio-economic,socialsupportandhealthvariables.
Wemodelledimpairmentintwoways:asindividualimpair-mentsandasatotalofdifferentimpairments(oneimpair-mentversusnoneandtwoormoreversusnone).
Weusedmultiplelinearregressionstodevelopamodelfortheeffectofimpairmentonwellbeing,carryingouttestsoftheeffectofimpairmentafteraddinginpotentialcon-foundingvariables.
Weexploredinteractionsbetweensocialsupport,specificimpairments,totalimpairmentsandtotaldisabilityinthemultivariablemodel.
AlltestswereWaldtestsasappropriateforweightedsurveydata.
Residualswerecomputedforthefinalmultivariablemodelandplottedashistograms(toassessanyevidencefornonnormality,includingindividualoutliers)andwerealsoplottedagainstpredictedvalues(toassessevi-denceforheteroscedasicity,inthesenseofgreaterspreadwithincreasingvalue).
Varianceinflationfactors(VIFs)werecomputedforallindependentvariablestocheckforcollinearity.
Results1620olderadultsin1620householdsweresampled,ofwhom1300(80%)wereeligibletotakepart.
Reasonsfornotbeingeligiblewerehavingnobiologicaloradoptedchildrenorstep-children;havingdiedsince2004,ormovedoutofthevillage.
Outthe1300eligible,1147(88%)agreedtotakepartand153(12%)werenon-respondersofwhom110wereunavailableforaninter-view(despiteatleastthreevisitstothehousehold),21refusedtotakepartand22weretoounwell.
Oftheresponders,datawereincompletefor43duetotheolderadultbeingunwellorcognitivelyimpaired.
Therewerenosignificantdifferencesbetweenrespondersandnon-respondersintermsofage,gender,livingalone,beingmarried,oreducation.
Demographicdescriptionofsample–Table1Table1showstheactualsamplenumbersandweightedestimateofthecharacteristicsinthewiderprovincepopu-lationofparentsfromwhichthesamplewasdrawn.
Theaverageagewas70years(SD7.
1).
AsshowninTable1,57%oftheparticipantswerefemale.
Nearlyhalfhadlessthanprimaryschooleducation,whichforoursamplemeantlessthanfouryearseducation.
(OnlyinthelasttwodecadeshasThailand'scompulsoryeducationextendedtosixandnowtotwelveyears)Nearlyhalfwerestillwork-ing.
Becauseweover-sampledthosenotco-residentwithachild,thestudypopulationhasalowerproportionliv-ingwithachildcomparedtotheprovinceestimateandisslightlymorelikelytolivealone.
Otherwisetherewerenegligibledifferencesbetweenthestudysampleandtheestimatedprovincepopulation.
Theaveragenumberoflivechildrenintheseparentswas4.
8(SD2.
4);2.
4sonsand2.
4daughters.
Three-quarterseitherlivedwithachildorsawachilddaily.
Themeandurationofresidenceinthesamedistrictwasnearly50years.
Themeanwellbeingscorewas33.
3(SD7.
6).
Associationbetweentypesofimpairmentsandwellbeing–Table2Thethreemostcommonimpairmentswerearthritis,pain,andeyesightproblems.
Approximatelyone-third(32%)oftheolderadultsdidnothaveanyimpairment,18%hadoneand50%hadtwoormoreimpairments.
Impairmentsduetoarthritis,pain,paralysis,vision,stomachproblemsorbreathingwereallassociatedwithloweredwellbeing.
Paralysis,faints/blackout,breathlessness,andpainweretheimpairmentswiththehighesteffectsizeforlesswell-being.
Afteradjustingtheimpairmentsfordisability,onlyparalysisremainedsignificantlyassociatedwithlowwell-being.
HealthandQualityofLifeOutcomes2009,7:66http://www.
hqlo.
com/content/7/1/66Page5of9(pagenumbernotforcitationpurposes)Associationbetweennumberofimpairmentsandwellbeing–Table3AsshowninTable3,havingoneimpairmentcomparedtononeandhavingtwoormorecomparedtononewassig-nificantlyassociatedwithlesswellbeing.
Thisassociationremainedafteradjustingforsocio-demographicfactors,socialsupportfromchildren,socialsupporttochildren,andsocialsupportfromothers.
Thereappearedtobesomepositiveconfoundingbysocio-demographicfactorsasthecoefficientsfortheassociationwithimpairmentfellslightlyandthestatisticalsignificancedecreased.
Thismaybeexplainedbecausefactorssuchaswealthandeducationareassociatedwithgreaterwellbeingandwithlessimpair-ment.
Thereappearedtobesomeslightnegativecon-foundingbysocialsupportfromandtochildrenasthesignificanceroseagainafteradjustingforthese.
Thiscouldbebecausemoreimpairedolderpeoplesarelikelytoreceivemoresocialsupportfromchildrenandothers,andmoresocialsupportisalsoassociatedwithgreaterwellbe-ing.
Finally,afteradjustingfordisability,theassociationbetweennumberofimpairmentsandwellbeingfellandwasnolongersignificant.
Multivariablemodel–Table4Variablesthatweresignificantlyassociatedwithwellbeingeitherbeforeand/orafteradjustmentareshowninTable4.
Theparsimoniousmultivariablemodelforpsychologi-calwellbeingincludedage,householdwealth,currentlyworking,familynetworksizeclose-by,receivingsupportfromchildren,receivingsupportfromothers,talkingmorefrequentlytoachild,perceivingreceivingveryade-quatesupportfromchildren,perceivinggivinggoodsup-porttochildren,lessimpairmentduetoparalysis,(p=0.
003),lessgeneralimpairment,lessdisability,andgivinglessactualsupporttochildren.
Ofnote,neitherlivingaloneorcognitiveimpairmentwereassociatedwithwell-Table1:Descriptivecharacteristicsofparents:actualsamplenumbers(totaln=1147)andweightedpercentagesStudysamplen=1147WeightedpercentagesFemalen=63457%Workingn=56448%Maritalstatus:Marriedn=63354%Widowedn=45141%Divorced/separated/singlen=636%Livealonen=1559%Education:Nonen=33228%1–3yearsn=17415%Primary(4yrs)n=54149%Morethanprimaryn=998%Proportionwithtwoormorelimitingphysicalimpairmentsn=54050%Cognitiveimpairmentn=918%Atleastonechildlivingathomen=55163%Table2:Prevalenceofimpairmentsandassociationswithwellbeing,weightedlinearregressionHealthimpairmentsWeightedpercentages(95%confidenceintervals)CoefficientforassociationwithwellbeingPvalueforassociationwithwellbeingPvalueforassociationwithwellbeing,adjustedfordisabilityArthritisorrheumatism44.
4(40.
0–48.
4)-1.
66<0.
0010.
915Eyesight23.
3(19.
3–27.
3)-2.
07<0.
0010.
202Hearing7.
6(6.
0–9.
2)-.
760.
4960.
843Cough3.
9(2.
4–5.
4)-2.
890.
1100.
306Breathing7.
7(5.
4–10.
0)-2.
730.
0240.
186Highbloodpressure16.
3(13.
0–19.
5)-0.
480.
4150.
185Diabetes7.
1(4.
8–8.
7)-1.
570.
2630.
788Hearttroubleorangina6.
4(4.
1–8.
7)-1.
120.
5340.
831Stomachorintestine9.
3(6.
6–12.
0)-2.
500.
0080.
086Faintsorblackouts17.
8(14.
5–20.
9)-2.
630.
0010.
143Paralysis2.
3(1.
1–3.
5)-4.
66<0.
0010.
012Skin3.
4(2.
2–4.
6)0.
020.
9930.
785Pain37.
3(32.
1–42.
4)-2.
46<0.
0010.
105HealthandQualityofLifeOutcomes2009,7:66http://www.
hqlo.
com/content/7/1/66Page6of9(pagenumbernotforcitationpurposes)being.
Thepercentageofvarianceexplainedbythemulti-variablemodelwas32%.
Theresidualsshowednoevidencefornonnormalitynorforoutliers,andtherewasnoevidenceforheteroscedascity.
Therewasnoevidenceforcollinearity(allVIFs<10).
Therewasaninteractionbetweensocialsupportfromchildrenandparalysis–thosewithlowreceivedsocialsupportfromchildrenandwithparalysiswereespeciallylikelytohavelowwellbeing(pvalueforinteraction0.
033).
DiscussionThekeyfindingfromthispaperisthatimpairmentduetoparalysiswasassociatedwithloweredpsychologicalwell-beinginolderThaipeople,evenaftercontrollingforelevenotherphysicalimpairments,disability,socio-eco-nomicfactorsandsocialsupport.
Asecondkeyfindingisthatwhileanincreasingnumberofimpairmentswasalsoassociatedwithlesswellbeing,thisassociation,andthosewithotherindividualimpairments,wereexplainedbydis-ability.
AthirdfindingisthatinthisThaisetting,receivedsupportfromadultchildoffspring,receivedsupportfromothersandperceivedsupportfromadultchildoffspringwereallindependentlyassociatedwithgreaterwellbeinginolderparentswhereasactualsupporttochildrenwasassociatedwithlowerwellbeing.
Chanceisanunlikelyexplanationfortheadjustedassoci-ationbetweenparalysisandlowwellbeing,andfortheadjustedassociationbetweendisabilityandlowwellbe-ing,astheassociationsweresignificantatalevelofp=0.
001.
Wewereabletoadjustforarangeofcovariatessoconfoundingisanunlikelyexplanation.
AllimpairmentTable3:Associationbetweenwellbeingscoreandhavingoneortwoormorephysicalimpairments(samplen=1147)NumberofphysicalimpairmentsCoefficientforhavingoneimpairmentcomparedtonone*Coefficientforhavingtwoormoreimpairmentscomparedtonone*WaldtestF(2,95)Pvalue-1.
55-3.
0315.
52<0.
001Adjustedforsocio-demographiccharacteristics1-1.
01-2.
558.
52<0.
001Adjustedfor1+socialsupportandsocialnetwork2-0.
64-2.
4313.
44<0.
001Adjustedfor1+2+socialsupporttochildren3-0.
53-2.
3714.
42<0.
001Adjustedfor1+2+3+disability4-0.
23-0.
480.
420.
656Adjustedfor1+2+3+4+cognitiveimpairment5-0.
21-0.
460.
380.
685Table4:Associationsbetweenpsychologicalwellbeinganddemographic,socialandphysicalhealthstatus(samplen=1147)UnadjustedCoefficientUnadjustedPvalueAdjustedcoefficient*AdjustedPvalue*OlderAge(years)0.
030.
4550.
130.
010Female-1.
250.
0270.
020.
980Currentlyworking0.
430.
4191.
320.
018Marriedversuswidowed/single/divorced1.
050.
0660.
330.
581Livealone-1.
060.
113-0.
360.
659Education1.
480.
0070.
150.
145Wealthyhousehold0.
89<0.
0010.
32<0.
001Physicalimpairment-0.
75<0.
001-.
370.
020Paralysis-4.
66<0.
001-2.
96<0.
001Disability-0.
29<0.
001-.
22<0.
001Cognitiveimpairment-1.
030.
218-1.
430.
223Familysocialnetworksize0.
11<0.
0010.
070.
002Atleastonechildlivinginhouseholdversusnochildreninthehousehold-0.
100.
850-0.
570.
304Talktoachildatleastweekly0.
930.
0020.
740.
029Receivingsupportfromchildren0.
51<0.
0013.
06<0.
001Receivingfinancialremittancesfromchildren2.
18<0.
0011.
55<0.
001Givingsupporttochildren0.
250.
284-0.
62<0.
001Receivingsupportfromothers0.
440.
0030.
530.
003Perceivegoodsupportfromchildren3.
79<0.
0013.
06<0.
001Perceivegivinggoodsupporttochildren3.
31<0.
0011.
260.
029*adjustedforallothervariablesinthetableinaweightedregression.
HealthandQualityofLifeOutcomes2009,7:66http://www.
hqlo.
com/content/7/1/66Page7of9(pagenumbernotforcitationpurposes)anddisabilitymeasuresreliedonsubjectiveperceptionwhichmayleadtomisclassificationofhealthstatus,althoughahighlevelofagreementhasbeenreportedbetweenself-reportedandobjectivehealthstatusmeas-ures[19].
Biasisunlikelyinthiscommunitysamplewithagoodresponserateandinterviewerswereblindtothestudyhypotheses.
Althoughweoversampled,thiswasonthebasisonlivingarrangementsratherthanhealthandwasanywaytakenaccountofintheanalysis.
Non-system-aticerrorispossible–forinstancethismighthavecomeaboutthroughpoorreliabilityoftheinterviewingteamorthroughparticipants'errorsinrecalloftheirhealthprob-lems,althoughpreviousworkhasshownahighlevelofagreementbetweenself-reportandobjectivehealthstatusmeasures[19].
Wedidnotformallyassessinter-raterreli-ability.
Howeveraspartofthedemographicsurveillancesystemapproach,qualitycontroliswellestablishedandprioritisedincludingdailychecksondatacompletenessandconsistency,havingaresearchsupervisorforeachteamofinterviewersandhavingfieldstationresearchmanagers(trainedintheinterviewbutblindtothehypothesis),andresearchers,infrequenttelephonecon-tactandmakingregularvisitstothedatacollectionteams.
Thisisacross-sectionalstudysothedirectionofcausalitycannotbedefinitelyinferred.
WhywasparalysisassociatedwithalargeandsignificanteffectonwellbeingStudiesofolderpeopleinWesterncountrieshavereportedlowmoodanddepressionpartic-ularlyfollowingstrokeandthatthisassociationwasinde-pendentofdisability[26].
Post-strokedepressionofcoursemayhaveabiologicalbasiswhichmayexplainourfinding[27].
However,wellbeingisabroaderconceptthandepression.
Ourmeasureofwellbeingwasdevel-opedandvalidatedusingthoroughqualitativeandquan-titativeworkwithThaiolderpeople[12,18]andincludesconceptsvitaltoThaiwellbeingincludinginterpersonalaswellasintrapersonalaspects.
TheeffectofparalysismaybeduetothescarcedisabilityservicesinruralThailand,withfewopportunitiestoreceiveaids,adaptations,orcommunitytransport.
Ruralpeoplemaythusbeespe-ciallyvulnerabletolossofsocialcontactsintheneigh-bourhoodandtolosingrespect.
Anotherpossibilityisthatimpactsofstrokegobeyonddisability,eitherviabiologi-caleffectsonthebrain[27]orthroughthepsychologicalmeaningofstrokesuchasshameoverlossoffunctionandalteredappearanceandfearsaboutprognosis.
Inthisset-tingofhighout-migration,absenceofchildrenmayalsobeafactor,althoughmostolderpeoplestilleitherliveclosetoachildortalktoachildweeklyormore.
Ourfindingthatdisabilityexplainstheassociationbetweennumberofimpairmentsandlowwellbeingech-oesstudiesthathavelookedatimpairment,disabilityanddepressionandatimpairmentsandwellbeinginWesterncountries[6,9,28,29].
Prospectivestudieshaveshownthatdisabilitycanpredicttheonsetofdepression[29].
Arecentreviewconcludedthatmuchoftheeffectofimpair-mentonnegativeaffectcouldbeexplainedbythepoten-tialmediatingeffectofdisability[30].
Itisstrikingthatourresultmirrorsthatfromwesterncountries,showingthecross-culturalapplicabilityofthewellbeingmodel.
Themodelforgreaterwellbeingincludedotherfactors,notablyreceivedsocialsupportfromchildren,perceivedsocialsupportfromchildren,receivedsocialsupportfromothers,financialremittancesfromchildrenandwealth.
Asanumberofassociationswereanalysedinthisstudy,aproblemofmultipletestingmighthaveoccurred.
How-ever,itisunlikelythatthiswouldexplainourfindingsasmostofthefactorsintheparsimoniousmodelforwellbe-ingweresignificantatp<0.
001orp=0.
001.
Severalpos-siblemechanismscouldexplaintheeffectofreceivedsocialsupportonwellbeing.
Socialsupportmayreducestressandconsequentlybuffertheeffectofnegativeevents.
Althoughreceivedsupportislikelytoreflectneed,certaintypesofreceivedsupportmaybevaluableinbring-ingaboutimprovedwellbeing[31].
Greatersocialsupportmightalsoaidolderpeoplewithimpairmenttocarryoutdailytasks,encouragethemtobephysicallyactive,increasemedicationcompliance,decreasesocialrestrictionandenhanceself-esteem[32].
IntheThaiculture,connectionsbetweenparentsandchil-drenarevital[33].
Althoughmanyparentsinthisstudyhadout-migrantchildren,theycontinuedtoreceivesup-portthroughtelephonecontact,visitsandeconomicremittances[17]Inaddition,theyreceivedsupportfromothers,oftenneighboursorotherrelativeslivingcloseby,andthiswasalsoindependentlyassociatedwithgreaterwellbeing.
Thissuggeststhatolderpeoplelivingwithoutchildrenareadaptingtotherealitiesofout-migrationandfindinghelpfromothersclosebyintheirneighbourhood.
Itisstrikingthatreceivedsupportfromchildrenandfromothersappearedhelpful,andthatreceivedsupportfromchildrenmayevenbuffertheimpactofparalysisonlowwellbeing.
OlderThaipeoplemayplacelessvalueonautonomythanthoseinwesterncountries,findingsup-portfromfamilymembersespeciallyimportantandcom-forting[12].
Aperceptionbytheparentofgivingagoodamountofsupporttotheiroffspringwasassociatedwithbetterwell-being.
However,givingactualsupporttochil-drenwasassociatedwithlesswellbeing,perhapsreflectingparentalobligationinthisculturetosupportadultchil-dreninneed[34].
Somelimitationsofthisstudyincludeitscross-sectionaldesign.
Secondlyourmeasureofwellbeingisculturespe-cific–althoughthismayalsoberegardedasstrengthofthestudy.
Thirdly,thefindingsfromthisstudymightlackHealthandQualityofLifeOutcomes2009,7:66http://www.
hqlo.
com/content/7/1/66Page8of9(pagenumbernotforcitationpurposes)generalisabilitytoallolderadultsasthesamplewasrestrictedtoparentswithatleastonelivingchild,althoughinThailandthisexcludedonly5%ofolderpeo-pleasweincludedanyonewithabiological,adoptedorstepchild.
Inconclusion,disabilitymaymediatemostoftheimpactofchronicphysicalimpairmentsonpsychologicalwellbe-ing,althoughparalysisappearstohaveanindependenteffect.
Receivedsocialsupport,perceivedsocialsupportandwealthalsohaveimportantpositiveeffectsonpsy-chologicalwellbeing.
Improvingdisabilityservicesandoptimisingsocialsupportwillbevitalinruralareasindevelopingcountrieswhicharelikelytoexperienceincreasingdepletionofyoungeradultsinthenextdecade.
Whilecareiscurrentlyprovidedbyfamilymembers,espe-ciallydaughtersandgrand-daughters,wesuggestthatpotentiallyvaluableservicesinruralareasmayincludehomecareprogrammesforolderpeopleandtheircarers,homevisitsbyhealthcarevolunteersinthevillage,daycare,extendingtheexistingnetworkof'elderlyclubs',occupationaltherapytoenableaidsandadaptationsathome,andmakingarangeoffacilitiesmoreaccessibletoolderdisabledpeople,ConclusionInconclusion,inthisThairuralsetting,mostoftheasso-ciationbetweenphysicalimpairmentsandlowerwellbe-inginolderpeopleisexplainedbydisability.
Receivedsupportfromchildrenandfromothersandperceivedhighsupportfromandtochildrenwereallindependentlyassociatedwithgreaterwellbeingwhereasgivingactualsupporttochildrenwasassociatedwithlowerwellbeing.
Improvingcommunitydisabilityservicesforolderpeopleandoptimizingreceivedsocialsupportthroughfamilies,neighboursandhomecareprogramswillbevitalinruralareasindevelopingcountries.
CompetinginterestsTheauthorsdeclarethattheyhavenocompetinginterests.
Authors'contributionsAllauthorsmadesubstantialcontributionstostudydesignandinterpretationofdata.
MAhadmainresponsi-bilityforanalysingdataanddraftingthemanuscript.
SPandKThadmainresponsibilityforacquisitionofdata.
Allauthorswereinvolvedinrevisingthemanuscriptcriticallyandhavegivenfinalapprovaloftheversiontobepub-lished.
AcknowledgementsWethankDrBenchaYoddumnern-Attig,DrPhilipGuestandProfMartinPrinceforadviceonthestudydesignandmethods,MsWanneeHutapatandMsJongjitRithirongfordatamanagement,DrRobertStewartforcom-mentsonthemanuscript,allthefieldstaff(NiphonDarawuttimaprakorn,JeerawanHongthong,PhattharaphonLuddakulWipapornJarruruengpaisanandYaowalakJiaranai)andparticipantsoftheKanchanaburiDemographicSurveillanceSystem,andtheWellcomeTrustforfundingtheproject(WT078567).
References1.
WorldHealthOrganisation:PromotingMentalHealth;Con-ceptsemergingevidenceandpractise,Summaryreport.
Geneva.
2004.
2.
WorldHealthOrganisation:Strengtheningmentalhealthpro-motion.
Geneva.
2001.
3.
RyffCD,KeyesCLM:Thestructureofpsychologicalwell-beingrevisited.
JournalofPersonalityandSocialPsychology1995,69(4):719-727.
4.
TennantR,HillerL,FishwickR,PlattS,JosephS,WeichS,ParkinsonJ,SeckerJ,Stewart-BrownS:TheWarwick-EdinburghMentalWell-beingScale(WEMWBS):developmentandUKvalida-tion.
HealthQualLifeOutcomes2007,5:63.
5.
ArakiA,MurotaniY,KamimiyaF,ItoH:LowWell-BeingIsanIndependentPredictorforStrokeinElderlyPatientswithDiabetesMellitus.
JournaloftheAmericanGeriatricsSociety2004,52(2):205-210.
6.
KendigH,BrowningCJ,YoungAE:Impactsofillnessanddisabil-ityonthewell-beingofolderpeople.
DisabilRehabil.
2000,22(1–2):15-22.
7.
SmithJ:Well-beingandhealthfromage70to100:findingsfromtheBerlinAgingStudy.
EuropeanReview2001,9(04):461-477.
8.
BeekmanAT,PenninxBW,DeegDJ,OrmelJ,BraamAW,vanTilburgW:Depressionandphysicalhealthinlaterlife:resultsfromtheLongitudinalAgingStudyAmsterdam(LASA).
JAffectDisord1997,46(3):219-231.
9.
OrmelJ,KempenGI,PenninxBW,BrilmanEI,BeekmanAT,vanSon-derenE:Chronicmedicalconditionsandmentalhealthinolderpeople:disabilityandpsychosocialresourcesmediatespecificmentalhealtheffects.
PsycholMed1997,27(5):1065-1077.
10.
PrinceMJ,HarwoodRH,BlizardRA,ThomasA,MannAH:Socialsupportdeficits,lonelinessandlifeeventsasriskfactorsfordepressioninoldage.
TheGospelOakProjectVI.
PsycholMed1997,27(2):323-332.
11.
KuPW,FoxK,McKennaJ:AssessingSubjectiveWell-beinginChineseOlderAdults:TheChineseAgingWellProfile.
SocialIndicatorsResearch2008,87(3):445-460.
12.
Ingersoll-DaytonB,SaengtienchaiC,KespichayawattanaJ,Aung-surochY:Measuringpsychologicalwell-being:insightsfromThaielders.
Gerontologist2004,44(5):596-604.
13.
SkeldonR:MigrationofWomenintheContextofGlobaliza-tionintheAsianandPacificRegion.
WomeninDevelopmentDis-cussionPaperSeries1999[http://www.
unescap.
org/esid/GAD/Publication/DiscussionPapers/02/series2.
pdf].
ESCAP14.
DeshingkarP:Internalmigration,povertyanddevelopmentinAsia.
2006[http://www.
odi.
org.
uk/resources/download/29.
pdf].
OverseasDevelopmentInstitute15.
ChouKL,ChiI:ReciprocalrelationshipbetweensocialsupportanddepressivesymptomsamongChineseelderly.
Aging&MentalHealth2003,7(3):224-231.
16.
Reportonbaselinesurveyround1.
NakhonPathom:InstituteforPopulationandSocialResearch,MahidolUniversity,2002;2000.
17.
AbasMA,PunpuingS,JirapramukpitakT,GuestP,TangchonlatipK,LeeseM,PrinceM:Rural-urbanmigrationanddepressioninageingfamilymembersleftbehind.
BritishJournalofPsychiatry2009,195:54-60.
18.
Ingersoll-DaytonB,SaengtienchaiC,KespichayawattanaJ,Aung-surochY:Psychologicalwell-beingAsianstyle:theperspec-tiveofThaielders.
JCrossCultGerontol2001,16(3):283-302.
19.
BurvillPW,MowryB,HallWD:Quantificationofphysicalillnessinpsychiatricresearchintheelderly.
InternationalJournalofGer-iatricPsychiatry1990,5:161-170.
20.
WengerGC:AlongitudinalstudyofchangesandadaptationinthesupportnetworksofWelshelderlyover75.
JournalofCross-CulturalGerontology1986,1(3):277-304.
21.
RindfussRR,JampaklayA,EntwisleB,SawangdeeY,FaustK,Pras-artkulP:TheCollectionandAnalysisofSocialNetworkDatainNangRong,Thailand.
InNetworkEpidemiology–AHandbookforPublishwithBioMedCentralandeveryscientistcanreadyourworkfreeofcharge"BioMedCentralwillbethemostsignificantdevelopmentfordisseminatingtheresultsofbiomedicalresearchinourlifetime.
"SirPaulNurse,CancerResearchUKYourresearchpaperswillbe:availablefreeofchargetotheentirebiomedicalcommunitypeerreviewedandpublishedimmediatelyuponacceptancecitedinPubMedandarchivedonPubMedCentralyours—youkeepthecopyrightSubmityourmanuscripthere:http://www.
biomedcentral.
com/info/publishing_adv.
aspBioMedcentralHealthandQualityofLifeOutcomes2009,7:66http://www.
hqlo.
com/content/7/1/66Page9of9(pagenumbernotforcitationpurposes)SurveyDesignandDataCollectionEditedby:MorrisM.
Oxford:OxfordUniversityPress;2004.
22.
SherbourneCD,StewartAL:TheMOSsocialsupportsurvey.
SocSciMed.
1991,32(6):705-714.
23.
GanguliM,ChandraV,GilbyJE,RatcliffG,SharmaSD,PandavR,Sea-bergEC,BelleS:Cognitivetestperformanceinacommunity-basednondementedelderlysampleinruralIndia:theIndo-U.
S.
Cross-NationalDementiaEpidemiologyStudy.
Interna-tionalPsychogeriatrics1996,8(4):507-524.
24.
WelshKA,ButtersN,MohsRC,BeeklyD,EdlandS,FillenbaumG,HeymanA:TheConsortiumtoEstablishaRegistryforAlzhe-imer'sDisease(CERAD).
PartV.
Anormativestudyoftheneuropsychologicalbattery.
Neurology1994,44(4):609-614.
25.
Epping-JordanJ,UstunT:TheWHODAS-II:levelingtheplayingfieldforalldisorders.
WHOMentalHealthBulletin2000,6:5.
26.
StewartR,PrinceM,MannA,RichardsM,BrayneC:Stroke,vascu-larriskfactorsanddepression:Cross-sectionalstudyinaUKCaribbean-bornpopulation.
BrJPsychiatry2001,178(1):23-28.
27.
KimJM,StewartR,KimSW,YangSJ,ShinIS,YoonJS:Vascularriskfactorsandincidentlate-lifedepressioninaKoreanpopula-tion.
TheBritishJournalofPsychiatry2006,189(1):26-30.
28.
BraamAW,PrinceMJ,BeekmanAT,DelespaulP,DeweyME,Geer-lingsSW,KivelaSL,LawlorBA,MagnussonH,MellerI,etal.
:PhysicalhealthanddepressivesymptomsinolderEuropeans.
ResultsfromEURODEP.
BrJPsychiatry2005,187:35-42.
29.
PrinceMJ,HarwoodRH,BlizardRA,ThomasA,MannAH:Impair-ment,disabilityandhandicapasriskfactorsfordepressioninoldage.
TheGospelOakProjectV.
PsycholMed1997,27(2):311-321.
30.
PrinceM,PatelV,SaxenaS,MajM,MaselkoJ,PhillipsMR,RahmanA:Nohealthwithoutmentalhealth.
Lancet2007,370(9590):859-877.
31.
Boutin-FosterC:Gettingtotheheartofsocialsupport:Aqual-itativeanalysisofthetypesofinstrumentalsupportthataremosthelpfulinmotivatingcardiacriskfactormodification.
Heart&Lung:TheJournalofAcuteandCriticalCare2005,34(1):22-29.
32.
GreenglassE,FiksenbaumL,EatonJ:Therelationshipbetweencoping,socialsupport,functionaldisabilityanddepressionintheelderly.
Anxiety,StressandCoping2006,19:15-31.
33.
KnodelJ,SaengtienchaiC,SittitraiW:ThelivingarrangementsofelderlyinThailand:viewsofthepopulace.
JournalofCross-Cul-turalGerontology1995,10:79-111.
34.
ZimmerZ,KorniekK,KnodelJ,ChayovanN:SupportbymigrantstotheirelderlyparentsinruralCambodiaandThailand:Acomparativestudy.
InPoverty,GenderandYouthWorkingPaperno2NewYork:PopulationCouncil;2007.